This was interesting to read, I especially liked the 4 quadrant comparison of formalism and decentralized vs centralized authority.
The article does answer the question 'why should managers exist' (to direct the subjective so the enterprise's limited means move towards its goals in the absence of enough AI smarts to know what is good right now).
I went into it expecting a 'rah rah fire all the managers' potboiler and was pleasantly surprised.
What I do wonder though is, businessman to businessman, how much of our economic activity really does operate in the academic 'free market' principles this article leans on...and I think that the answer to that question directly drives the application of this article's follow-up question at least to my mind: "will AI eat manager jobs too?" and the answer is murky even to me.
I think reputation, ego, feelings/sales, habits and circumstances have unaccounted for influences on the free market, to say nothing of the 'businesses want to capture the market, government wants to regulate the market, consumers want to eat the market' bigger non-academic forces at play we've all discussed so often in the past.
Management has come a long way since the railroad. Management (as a task not just a role) is more multi-faceted than considered in the article.
The emphasis here is on authority, but to me accountability is the more interesting question. How can one trust the integration? That's not a simple validation question once you cross a big enough boundary.
The "negotiating agents" take is interesting. But I feel like we're in the "I'm feeling lucky" stage of LLMs. I do not go to Google search results today and "feel lucky." Adversarial incentives will do the same thing to LLM negotiations.
This was interesting to read, I especially liked the 4 quadrant comparison of formalism and decentralized vs centralized authority.
The article does answer the question 'why should managers exist' (to direct the subjective so the enterprise's limited means move towards its goals in the absence of enough AI smarts to know what is good right now).
I went into it expecting a 'rah rah fire all the managers' potboiler and was pleasantly surprised.
What I do wonder though is, businessman to businessman, how much of our economic activity really does operate in the academic 'free market' principles this article leans on...and I think that the answer to that question directly drives the application of this article's follow-up question at least to my mind: "will AI eat manager jobs too?" and the answer is murky even to me.
I think reputation, ego, feelings/sales, habits and circumstances have unaccounted for influences on the free market, to say nothing of the 'businesses want to capture the market, government wants to regulate the market, consumers want to eat the market' bigger non-academic forces at play we've all discussed so often in the past.
If we fired all the managers, what would I do???
Management has come a long way since the railroad. Management (as a task not just a role) is more multi-faceted than considered in the article.
The emphasis here is on authority, but to me accountability is the more interesting question. How can one trust the integration? That's not a simple validation question once you cross a big enough boundary.
The "negotiating agents" take is interesting. But I feel like we're in the "I'm feeling lucky" stage of LLMs. I do not go to Google search results today and "feel lucky." Adversarial incentives will do the same thing to LLM negotiations.
History of managers starting with railroads and how more intelligence has been added to what managers do over time